What Is the Labor Share of Income? Formula and Measurement

A labor share asks how much of production income is attributed to labor under a stated accounting definition. In a value-added series, the basic formula is:

Labor share = labor compensation ÷ nominal value added × 100

“Nominal” means that both amounts are valued at the prices of the same period. If a sector produces $500 billion of gross value added and its labor compensation is $300 billion, its labor share under that definition is 60%. The ratio is a share of an aggregate flow over a period, not a count of jobs, a wage rate, or the fraction of a paycheck a worker keeps.

There is not one universal series called “the labor share.” A published statistic may describe all domestic production, private business, nonfarm business, manufacturing, or a particular industry. It may use value added or another output boundary, include only employees in the numerator, or estimate a labor component for proprietors. Name the series before comparing its level or movement. The OECD’s productivity compendium, for example, defines its labor-share indicator as total labor compensation divided by gross value added within the stated coverage. {source:oecdLaborIncomeProductivity2024}

Compensation is broader than a worker’s cash wage. In the national accounts it generally includes wages and salaries plus supplements paid by employers, such as contributions to pensions, insurance, and government social-insurance programs. These payments are treated as labor compensation even when they are not handed to employees as cash in that pay period. {source:beaEmployeeCompensation2023}

That definition makes the measure useful for studying the cost and income attributed to labor as a production input, but it differs from take-home pay. Taxes withheld from pay, household transfers, and the timing of a worker’s paycheck are not the same concept as compensation in the production accounts. Nor does an aggregate compensation total say how much a typical worker receives: it combines workers with different wages, hours, benefits, occupations, and industries.

For a small hypothetical sector, suppose value added is $100, employees receive $50 in cash wages, and the employer pays $10 toward benefits. Employee compensation is $60, so the compensation-based labor share is 60%. Dividing wages alone by output would give 50%, but that would use a narrower numerator. The $60 is not the employees’ after-tax take-home pay; it includes employer payments made on their behalf.

Statistical programs do not all build the numerator the same way. BEA’s national accounts classify employee compensation separately from proprietors’ income. By contrast, BLS labor-share estimates for some sectors impute a labor-compensation component for proprietors using assumptions about their work and comparable employee pay. The headline term can therefore hide a consequential population and estimation choice. {source:beaEmployeeCompensation2023} {source:blsLaborShareEstimates2017}

For a value-added labor share, the denominator is the value of output after subtracting intermediate goods and services used up in production. This boundary avoids treating every business-to-business purchase as new value created by the measured sector. BLS uses labor compensation divided by value-added output for its value-added labor-share measure. Its handbook also describes measures based on sectoral production, where the output and cost boundary differs. Those ratios should not be read as interchangeable just because each is labeled a labor share. {source:blsLaborShareCalculation}

Gross and net value added answer another measurement question. Net value added subtracts consumption of fixed capital—the accounting estimate of capital assets used up during the period—from gross value added. If the same compensation is divided by a smaller net-value-added denominator, the calculated share is higher. That change does not mean workers suddenly received more compensation; the denominator excludes a different component of production income. BLS’s methodological discussion explains why gross and net output can be useful for different comparisons. {source:blsLaborShareEstimates2017}

Also check whether a statistic is expressed against gross domestic product, gross value added, national income, or an industry-specific measure. They are related national-account aggregates, but their boundaries and components differ. A ratio over GDP is not automatically the same as a ratio over gross value added: GDP at market prices includes taxes on production and imports less subsidies, while a value-added series can use a different production boundary. BEA’s income-side GDP breakdown also lists gross operating surplus and consumption of fixed capital alongside employee compensation and net production taxes, so the complement of compensation is not a pure profit measure. {source:beaIncomeApproach} Match the numerator, denominator, economy, sector, and price basis before treating two percentages as a time series.

Suppose a hypothetical sector produces $100 of gross value added and reports $60 of employee compensation. The unadjusted employee-compensation share is $60 ÷ $100 = 60%. If the statistical series instead uses net value added after subtracting $10 of consumption of fixed capital, the same $60 numerator yields $60 ÷ $90 = 66.7%. The second percentage is larger because the denominator changed, not because the employees received an additional $6.70.

Now suppose an analyst estimates $8 of labor income for proprietors whose business income combines their work with returns to business assets. If that imputed amount is added to the numerator while gross value added remains $100, the adjusted share becomes ($60 + $8) ÷ $100 = 68%. This is another invented illustration, not an observation. It shows how a stated imputation can change the series. It does not establish that the proprietors actually earned exactly $8 for their labor, and it must not be described as extra employee wages.

The three figures—60%, 66.7%, and 68%—answer different questions because the numerator or denominator differs. A useful comparison records the definition next to the number rather than presenting a percentage without its accounting context.

A salaried employee’s compensation can be recorded separately from a corporation’s operating surplus. A sole proprietor’s business income is harder to split: it may compensate the owner’s work while also reflecting the use of the owner’s capital, risk, and entrepreneurship. In the U.S. NIPAs, nonfarm proprietors’ income includes these returns together rather than identifying a directly observed labor-only amount. {source:beaProprietorsIncome2024}

To estimate a labor share that includes owner-operators, statisticians may impute their labor earnings. One method assigns self-employed people the average hourly compensation of comparable employees and multiplies that rate by hours worked. The assumption is practical, not an observation of each proprietor’s wage. An owner may work longer or shorter hours, earn a different return for the same hours, or receive income from assets as well as labor. OECD research compares alternative imputation assumptions and explains why the adjusted result remains method-dependent. {source:blsLaborShareEstimates2017} {source:oecdSelfEmployedLaborShare2011}

This is why two reputable datasets can show different labor-share levels without either being an arithmetic error. One may count employee compensation only; another may add an estimated owner-labor component. They may cover different sectors or use a different value-added concept as well. For international comparisons, check how each statistical system classifies mixed income and whether the adjustments are harmonized. A decimal point does not remove those definitional choices.

An economy-wide labor share is a weighted aggregate of sectors whose shares can differ. It can move when the mix of output shifts, even if each sector’s own ratio stays unchanged. Consider a purely hypothetical economy with two sectors. Sector A produces $100 of value added and pays $70 in compensation, for a 70% share. Sector B also produces $100 but pays $30, for a 30% share. Together, compensation is $100 out of $200 of value added, or 50%.

Suppose sector A’s value added doubles to $200 while its 70% share stays the same; sector B remains at $100 and its 30% share also stays the same. The new aggregate is ($140 + $30) ÷ ($200 + $100) = 56.7%. No within-sector share changed. The economy-wide ratio rose because a larger portion of measured value added came from the higher-share sector. These invented values illustrate aggregation, not a claim about any country or a forecast.

In actual data, sector weights, relative prices, outsourcing, business organization, and revisions can all affect an aggregate measure. A time-series change therefore needs to be decomposed before it is attributed to a change in workers’ bargaining power, technology, or any other single cause. The ratio alone cannot identify a mechanism. The BLS study discusses the scope and estimation choices behind its U.S. labor-share series; its historical estimates should not be mistaken for current values or universal sector behavior. {source:blsLaborShareEstimates2017}

Labor share can summarize labor compensation relative to a clearly defined output measure. It can help analyze how income generated in production is divided between the compensation recorded for labor and the remaining accounting components. It can also contribute to a comparison between productivity growth and compensation growth when the price indexes, hours, and production boundary match. {source:blsLaborShareCalculation} {source:oecdLaborIncomeProductivity2024}

It does not show how income is distributed among workers. An economy could have a stable aggregate labor share while compensation shifts from lower-paid workers toward higher-paid managers. A falling share does not mean that every worker’s nominal pay or purchasing power fell; employment, hours, benefits, prices, and the distribution of compensation may move differently. Use wage, median-income, employment, and household-distribution measures for those questions.

For example, two hypothetical economies each produce $1,000 of value added and record $600 in total labor compensation, so both have a 60% labor share. In one, 10 workers each receive $60. In the other, nine receive $40 and one receives $240. Total compensation and the aggregate share are unchanged, but the second economy has a lower median. The invented figures show why an aggregate share cannot substitute for a distribution measure.

The remainder after labor compensation is deducted depends on the accounting boundary. BEA’s income-side GDP breakdown lists employee compensation alongside gross operating surplus, consumption of fixed capital, and taxes on production and imports less subsidies. The complement of a labor-share ratio can therefore include several components; the ratio alone does not establish a corporate-profit share. {source:beaIncomeApproach} {source:blsLaborShareCalculation}

Labor share also differs from unit labor cost. Unit labor cost relates compensation to real output, or hourly compensation to labor productivity, and is used to examine labor cost per unit produced. Labor share divides compensation by a current-price output measure. They are connected by accounting relationships but answer different questions and can move differently when output prices or definitions change. For a productivity–pay decomposition, see why productivity can outpace real compensation. For labor costs per unit, see unit labor costs and inflation. For domestic production versus resident income, see GDP, GNI, and GNP.

Before quoting a percentage, write down six details: the country or economy, the sector covered, the compensation numerator, the output denominator, whether values are gross or net, and how self-employed labor is treated. Then confirm the time period, current-price basis, and data vintage. If two reports differ, compare these definitions first; only then ask whether their methods or estimates disagree.

Keep interpretation proportional to the evidence. “This series records compensation equal to 60% of gross value added in the specified sector” is a precise statement. “Workers received 60% of the economy’s income” may not be, if the series excludes proprietors, covers only one sector, or uses a different output boundary. “The remainder went to corporations as profit” is not supported by the ratio alone.

A labor-share measure is most useful when its accounting choices are visible. Pair it with productivity for production per hour, real compensation for purchasing power, and income-distribution data for differences among households. No one ratio answers all of those questions.